FP64 FP64 Tensor Core FP32 34 TFLOPS 67 TFLOPS 67 TFLOPS H200 NVL1 34 TFLOPS 67 TFLOPS 67 TFLOPS -------- ------------- TF32 Tensor Core2 BFLOAT16 Tensor Core2 FPl 6 Tensor Core2 FPS Tensor Core2 INTS Tensor Core2 GPU Memory GPU Memory Bandwidth Decoders Confidential Computing Max Thermal Design Power (TDP) Multi-Instance GPUs Form Factor Interconnect Server Options NVIDIA Al Enterprise 1. Preliminary specifications. May be subject to change. 2. With sparsity. 989 TFLOPS 1,979 TFLOPS 1,979 TFLOPS 3,958 TFLOPS 3,958 TFLOPS 141GB 4.8TB/s 7 NVDEC 7 JPEG Supported Up to 700W (configurable) Up to 7 M IGs @l 8GB each SXM NVIDIA NVLink'": 900GB/s PCle Gens: l 28GB/s NVIDIA HGX'" H200 partner and NVIDIACertified Systems'" with 4 or 8 GPUs Add-on Ready to Get Started? To learn more about the NVIDIA H200 NVL Tensor Core GPU, visit www.pny.com/nvidia-h200nvl © 2024 NVIDIA Corporation and affiliates. All rights reserved. NVIDIA, the NVIDIA logo, HGX, Hopper, MGX, NVIDIACertified Systems, and NVLink are trademarks and/or registered trademarks of NVIDIA Corporation and affiliates in the U.S. and other countries. Other company and product names may be trademar·ks of the respective owners with which they are associated. 3418659. AUG24 989 TFLOPS 1,979 TFLOPS 1,979 TFLOPS 3,958 TFLOPS 3,958 TFLOPS 141GB 4.8TB/s 7 NVDEC 7 JPEG Supported Up to 600W (configurable) Up to 7 MIGs @l 8GB each PCle 2- or 4-way NVIDIA NVLink bridge: 900GB/s PCle Gens: l 28GB/s NVIDIA MGX"' H200 NVL partner and NVIDIA-Certified Systems with up to 8 GPUs Included H200SXM1 T